Every LLM has a default voice and it's making us all sound the same

A viral observation on Reddit and Discord has highlighted a pervasive issue in AI writing: despite using different prompts and models, outputs from tools like ChatGPT, Claude, and Llama often converge on a single, generic "AI voice." This voice is characterized by excessive formality, predictable structure, and a lack of authentic human variance, making content from different people sound eerily similar. The phenomenon suggests that large language models (LLMs) have a strong default stylistic setting that overrides user intent, homogenizing digital communication.

In response, a new tool called Noren is being developed to tackle this monotony head-on. Instead of generating text immediately, Noren's core innovation is to first analyze and learn a user's personal writing style—including their tone, sentence structure, and vocabulary preferences. By building a unique profile, the AI can then tailor its outputs to match the individual's authentic voice, aiming to break the cycle of sanitized, uniform content. The project, which is still in its early stages, represents a shift from one-size-fits-all generation to personalized AI assistance that preserves stylistic diversity.